VoC Analyses Part 2: Regressions

Participant info

age_group	N	meanAge	sdAge	nFemale
Children	17	11.32	0.8799	8
Adolescents	29	15.37	1.518	14
Adults	46	22.11	2.36	25

Agency task: Machine selection

Model: Optimal machine choices across trials by condition and age

## Mixed Model Anova Table (Type 3 tests, LRT-method)
## 
## Model: stage_2_acc ~ condition * zTrialOfCond * (zAge) + (condition * 
## Model:     zTrialOfCond | subID)
## Data: banditTask.filtered
## Df full model: 18
##                        Effect df     Chisq p.value
## 1                   condition  1 20.86 ***   <.001
## 2                zTrialOfCond  1 60.59 ***   <.001
## 3                        zAge  1      1.46    .227
## 4      condition:zTrialOfCond  1      0.08    .781
## 5              condition:zAge  1      0.21    .650
## 6           zTrialOfCond:zAge  1      1.46    .226
## 7 condition:zTrialOfCond:zAge  1      0.09    .770
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1

## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: stage_2_acc ~ condition * zTrialOfCond * (zAge) + (condition *  
##     zTrialOfCond | subID)
##    Data: data
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06))
## 
##      AIC      BIC   logLik deviance df.resid 
##   8674.1   8807.6  -4319.1   8638.1    12262 
## 
## Scaled residuals: 
##      Min       1Q   Median       3Q      Max 
## -16.3075   0.0788   0.1984   0.3889   2.4780 
## 
## Random effects:
##  Groups Name                    Variance Std.Dev. Corr             
##  subID  (Intercept)             2.4576   1.5677                    
##         condition1              0.3558   0.5965   -0.09            
##         zTrialOfCond            0.4269   0.6534    0.77  0.01      
##         condition1:zTrialOfCond 0.2272   0.4766    0.18  0.36  0.13
## Number of obs: 12280, groups:  subID, 92
## 
## Fixed effects:
##                              Estimate Std. Error z value Pr(>|z|)    
## (Intercept)                   2.37787    0.17242  13.792  < 2e-16 ***
## condition1                   -0.41517    0.08149  -5.095  3.5e-07 ***
## zTrialOfCond                  0.71264    0.08315   8.571  < 2e-16 ***
## zAge                          0.20570    0.16936   1.215    0.225    
## condition1:zTrialOfCond      -0.01845    0.06823  -0.270    0.787    
## condition1:zAge              -0.03405    0.07434  -0.458    0.647    
## zTrialOfCond:zAge             0.09524    0.07818   1.218    0.223    
## condition1:zTrialOfCond:zAge  0.01863    0.06182   0.301    0.763    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) cndtn1 zTrlOC zAge   cn1:TOC cnd1:A zTOC:A
## condition1  -0.108                                           
## zTrialOfCnd  0.713 -0.045                                    
## zAge         0.021 -0.003  0.033                             
## cndtn1:zTOC  0.096  0.473 -0.007 -0.001                      
## condtn1:zAg -0.003  0.091 -0.002 -0.096  0.079               
## zTrlOfCnd:A  0.034 -0.003  0.067  0.715 -0.009  -0.014       
## cndt1:TOC:A -0.002  0.078 -0.009  0.131  0.091   0.397  0.044

Plot: Proportion optimal machine selections across age groups and trials

Agency task: Agency decisions

Model: Agency decisions by VoC

## Mixed Model Anova Table (Type 3 tests, LRT-method)
## 
## Model: agency ~ zVoC * zTrialOfCond * zAge + (zVoC * zTrialOfCond | 
## Model:     subID)
## Data: banditTask
## Df full model: 18
##                   Effect df      Chisq p.value
## 1                   zVoC  1 144.39 ***   <.001
## 2           zTrialOfCond  1     3.56 +    .059
## 3                   zAge  1       0.01    .906
## 4      zVoC:zTrialOfCond  1  50.21 ***   <.001
## 5              zVoC:zAge  1     4.17 *    .041
## 6      zTrialOfCond:zAge  1       2.53    .111
## 7 zVoC:zTrialOfCond:zAge  1    9.59 **    .002
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1

## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: agency ~ zVoC * zTrialOfCond * zAge + (zVoC * zTrialOfCond |  
##     subID)
##    Data: data
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06))
## 
##      AIC      BIC   logLik deviance df.resid 
##  26597.3  26746.3 -13280.7  26561.3    28962 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -33.889  -0.549   0.190   0.543  13.837 
## 
## Random effects:
##  Groups Name              Variance Std.Dev. Corr             
##  subID  (Intercept)       1.86467  1.3655                    
##         zVoC              0.45871  0.6773   -0.18            
##         zTrialOfCond      0.23503  0.4848    0.63 -0.02      
##         zVoC:zTrialOfCond 0.09566  0.3093   -0.02  0.47 -0.34
## Number of obs: 28980, groups:  subID, 92
## 
## Fixed effects:
##                        Estimate Std. Error z value Pr(>|z|)    
## (Intercept)             0.74203    0.14391   5.156 2.52e-07 ***
## zVoC                    1.42145    0.07444  19.096  < 2e-16 ***
## zTrialOfCond            0.10238    0.05418   1.890  0.05879 .  
## zAge                   -0.01696    0.14368  -0.118  0.90603    
## zVoC:zTrialOfCond       0.32578    0.03909   8.335  < 2e-16 ***
## zVoC:zAge               0.15339    0.07390   2.076  0.03792 *  
## zTrialOfCond:zAge      -0.08631    0.05364  -1.609  0.10759    
## zVoC:zTrialOfCond:zAge  0.12286    0.03816   3.219  0.00129 ** 
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##             (Intr) zVoC   zTrlOC zAge   zVC:zTOC zVC:zA zTOC:A
## zVoC        -0.157                                            
## zTrialOfCnd  0.603 -0.002                                     
## zAge         0.003  0.000  0.001                              
## zVC:zTrlOfC -0.005  0.431 -0.219  0.001                       
## zVoC:zAge    0.000  0.015  0.002 -0.161  0.013                
## zTrlOfCnd:A  0.001  0.001  0.012  0.604 -0.005   -0.006       
## zVC:zTrOC:A  0.001  0.013 -0.004 -0.008  0.041    0.427 -0.242

Model: Agency decisions when VoC = 0

## Mixed Model Anova Table (Type 3 tests, LRT-method)
## 
## Model: agency ~ zAge + (1 | subID)
## Data: banditTask.vocZeroTrials
## Df full model: 3
##   Effect df Chisq p.value
## 1   zAge  1  0.00    .961
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1

## Generalized linear mixed model fit by maximum likelihood (Laplace
##   Approximation) [glmerMod]
##  Family: binomial  ( logit )
## Formula: agency ~ zAge + (1 | subID)
##    Data: data
## Control: glmerControl(optimizer = "bobyqa", optCtrl = list(maxfun = 1e+06))
## 
##      AIC      BIC   logLik deviance df.resid 
##   4304.7   4323.7  -2149.4   4298.7     4137 
## 
## Scaled residuals: 
##     Min      1Q  Median      3Q     Max 
## -4.2769 -0.6929  0.3759  0.6286  2.1169 
## 
## Random effects:
##  Groups Name        Variance Std.Dev.
##  subID  (Intercept) 1.24     1.113   
## Number of obs: 4140, groups:  subID, 92
## 
## Fixed effects:
##              Estimate Std. Error z value Pr(>|z|)    
## (Intercept)  1.246212   0.124404  10.017   <2e-16 ***
## zAge        -0.006025   0.123891  -0.049    0.961    
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '.' 0.1 ' ' 1
## 
## Correlation of Fixed Effects:
##      (Intr)
## zAge -0.002

Plot: Sensitivity to the value of choice

Plot: Sensitivity to value of control with continuous age

Summary stats: Sensitivity to value of control

Choice preference task

Choice preference task: summary stats

Model: Choice preference task accuracy

## Mixed Model Anova Table (Type 3 tests, LRT-method)
## 
## Model: correct ~ zDiff * zAge + (zDiff | subID)
## Data: rewardSense.filtered
## Df full model: 7
##       Effect df     Chisq p.value
## 1      zDiff  1 79.85 ***   <.001
## 2       zAge  1      0.78    .376
## 3 zDiff:zAge  1      1.36    .244
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1

Explicit reward knowledge task

Explicit reward knowledge task: summary stats

Model: Explicit reward knowledge by age and true probabilities

## Mixed Model Anova Table (Type 3 tests, S-method)
## 
## Model: error ~ zTrueProb * zAge + (1 | subID)
## Data: explicitKnow.filtered
##           Effect        df         F p.value
## 1      zTrueProb 1, 456.86 21.67 ***   <.001
## 2           zAge  1, 89.88      0.04    .845
## 3 zTrueProb:zAge 1, 456.86      0.36    .548
## ---
## Signif. codes:  0 '***' 0.001 '**' 0.01 '*' 0.05 '+' 0.1 ' ' 1

Plot: Explicit reward knowledge